Autism encompasses a striking heterogeneity of behaviors, which has so far precluded development of biologically meaningful measures of severity. The lack of reliable symptom clusters or behavioral dimensions along which to characterize individuals is a major obstacle to research into the biological causes, prognosis, and treatment of autism. Put simply, without a severity scale, individuals can never be shown to improve. Beyond this, the period during which autistic symptoms evolve and are probably most amenable to treatment is markedly understudied. Despite the dramatic cognitive and emotional changes during childhood and adolescence, little is known about the brain changes that underlie them. A number of studies have uncovered abnormalities in the emotional circuitry of the brain in autism, but results are inconsistent and have not been reliably linked to emotional and social behavior. The amygdala is known to be essential to fear behavior and to normal processing of facial expressions of emotion, but its role in autism is highly controversial. Using the tools of affective neuroscience and new magnetic resonance imaging techniques, we will focus on objective, quantitative measures of social and emotional deficits to characterize differences in amygdala structure, connectivity and biochemistry between individuals with autism. Our initial studies in this area showed amygdala volume to be strongly tied to these behavioral measures and to clinical diagnostic measures from early childhood. Importantly, this relationship was specific to nonverbal social and communicative impairments in autism. Our results were consistent with a model of chronic amygdala hyperactivity in autism that could resolve much of the inconsistency in previous analyses. This work will directly test several hypotheses from this model, including a longitudinal assessment of amygdala growth and the direct measurement of neurotransmitter levels in vivo. This work, focused on adolescence, will characterize the development of brain systems involved in emotion and affect and their dysregulation in autism. By tying autistic social impairments to specific brain differences, it will provide tools for future genetics studies, objective measures to gauge the efficacy of therapies, and knowledge of brain chemical abnormalities that will guide future pharmacological research.